EI Compendex Source List（2020年1月）
EI Compendex Source List（2019年5月）
EI Compendex Source List（2018年9月）
EI Compendex Source List（2018年5月）
EI Compendex Source List（2018年1月）
中国科学引文数据库来源期刊列
CSSCI(2017-2018)及扩展期刊目录
2017年4月7日EI检索目录（最新）
2017年3月EI检索目录
最新公布北大中文核心期刊目录
SCI期刊（含影响因子）
EI Compendex Source List
论文范文
1. Introduction Big data [1] has been sharply in progress unprecedentedly in recent years and is changing the operation for business as well as the decision-making for the enterprise. The huge amounts of data contain valuable information, such as the growth trend of system application and the correlation among systems. The undisclosed information may contain unknown knowledge and application that are discoverable further. However, big data with the features of high volume, high velocity, and high variety as well as in face of expanding incredible amounts of data, several issues emerging in big data such as storage, backup [2], management, processing, search [3], analytics, practical application, and other abilities to deal with the data also face new challenges. Unfortunately, those cannot be solved with traditional methods and thus it is worthwhile for us to continue exploring how to extract the valuable information from the huge amounts of data. According to the latest survey reported from American CIO magazine, 70% of IT operation has been done by batch processing in the business, which makes it “unable to control processing resources for operation as well as loading” [4]. This becomes one of the biggest challenges for big data application. Hadoop distributes massive data collections across multiple nodes, enabling big data processing and analytics far more effectively than was possible previously. Spark, on the other hand, does not do distributed storage [5]. It is nothing but a data processing tool, operating on those distributed data collections. Furthermore, Hadoop includes not only a storage component called Hadoop Distributed File System (HDFS), but also a processing component called MapReduce. Spark does not come with its own file management system. Accordingly, it needs to be integrated with Hadoop to share HDFS. Hadoop processing mostly static and batch-mode style can be just fine and originally was designed to handle crawling and searching billions of web pages and collecting their information into a database [6]. If you need to do analytics on streaming data, or to run required multiple operations, Spark is suitable for those. As a matter of fact, Spark was designed for Hadoop; therefore, data scientists all agree they are better together for a variety of big data applications in the real world. Through establishing a set of multiple big data analytics platforms with high efficiency, high availability, and high scalability [7], this paper aims to integrate different big data platforms to achieve the compatibility with any existing business intelligence (BI) [8] together with related analytics tools so that the enterprise needs not change large amounts of software for such platforms. Therefore, the goal of this paper is to design the optimization for job scheduling using MSHEFT algorithm as well as to implement optimized platform selection, and established platforms support R command to execute data retrieval and data analytics in big data environment. In such a way the upper-level tools relying on relational database which has stored the original data can run on the introduced platforms through minor modification or even no modification to gain the advantages of high efficiency, high availability, and high scalability. I/O delay time can be shared through reliable distributed file system to speed-up the reading of a large amount of data. Data retrieval and data analytics stack has layered as shown in Figure 1. As a result, according to performance index calculated for various methods, we are able to check out whether or not the proposed approach can reduce the execution time for the data retrieval and analytics significantly. 
|
|
|
